Alibaba Cloud·Qwen

Qwen2-72B

The largest dense Qwen2 model, using GQA across the family and extending support to 27 additional languages.

LegacyLanguageCodingOpen weights

MODEL NUMBER021

Specifications

Only published specifications are shown.

Total parameters

72.7B

Architecture

Dense

Context

128K

Input modalities

Text

Open weights

Yes

License

Apache-2.0

Status

Legacy

The largest dense Qwen2 model, using GQA across the family and extending support to 27 additional languages.

Qwen2 combined permissive Apache 2.0 licensing with strong multilingual capability, widening deployment options.

Compared with Qwen-7B ↗

PreviousQwen-7B7.7B · 32K ctx

CurrentQwen2-72B72.7B · 128K ctx

New capabilities Coding Math

GQA

MultilingualCodingMath

2024-06-07Open weights

The open Qwen2 family launched with a 72B flagship.

Select a card to view the related model.